Lagged associations between ambient temperature and first-ever ischaemic stroke onset by smoking status: a distributed lag non-linear model study

Ambient temperature may influence ischaemic stroke occurrence, but whether smoking modifies this association remains unclear. We examined lagged associations between six temperature indicators and first-ever ischaemic stroke and assessed effect modification by smoking status. We conducted a time-series study in Changchun, Northeast China. After excluding 2022 records to reduce the potential influence of COVID-19-related changes in healthcare utilisation, 2,425 patients with first-ever ischaemic stroke were included from 1 January 2023 to 7 April 2024, including 1,205 smokers and 1,220 non-smokers. Daily case counts were defined by symptom-onset date. Associations with average, apparent, minimum and maximum temperatures, diurnal temperature range, and temperature change between neighbouring days were estimated using quasi-Poisson generalised linear models combined with distributed lag non-linear models. Models were adjusted for relative humidity, temporal trends, seasonality, day of the week, and public holidays. Effect modification was tested using temperature-by-smoking interaction terms. Exploratory laboratory analyses applied false discovery rate correction. Lower average, apparent, minimum, and maximum temperatures were associated with increased daily case counts after a delay of approximately 2–3 weeks. Higher temperatures showed inverse associations at selected lags. Lower diurnal temperature range was associated with increased case counts mainly at lag 0. Temperature change between neighbouring days showed weaker, direction-dependent associations primarily within the first 2 weeks: warming was positively associated with case counts, whereas cooling was inversely associated. Formal interaction tests found no evidence of effect modification by smoking status (unadjusted P = 0.314–0.623; all Holm-adjusted P = 1.00). Among 1,688 patients with complete laboratory data, some indicators differed across absolute-temperature categories after false discovery rate correction in the overall and non-smoking groups, but not in the smoking group. Lower absolute temperatures were associated with increased first-ever ischaemic stroke case counts approximately 2–3 weeks later, whereas lower diurnal temperature range showed mainly a same-day association. Temperature changes between neighbouring days showed direction-dependent short-term associations. No statistically significant temperature-by-smoking interaction was identified. These findings support sustained cold-season preparedness and post-exposure health monitoring alongside smoking cessation and vascular risk management.